Fast Reconstruction of 3D Density Distribution around the Sun Based on the MAS by Deep Learning

Abstract

This study is the first attempt to generate a three-dimensional (3D) coronal electron density distribution based on the pix2pixHD model, whose computing time is much shorter than that of the magnetohydrodynamic (MHD) simulation. For this, we consider photospheric solar magnetic fields as input, and electron density distribution simulated with the MHD Algorithm outside a Sphere (MAS) at a given solar radius is taken as output. We consider 155 pairs of Carrington rotations as inputs and outputs from 2010 June to 2022 April for training and testing. We train 152 deep-learning models for 152 solar radii, which are taken up to 30 solar radii. The artificial intelligence (AI) generated 3D electron densities from this study are quite consistent with the simulated ones from lower radii to higher radii, with an average correlation coefficient 0.97. The computing time of testing data sets up to 30 solar radii of 152 deep-learning models is about 45.2 s using the NVIDIA TITAN XP graphics-processing unit, which is much less than the typical simulation time of MAS. We find that the synthetic coronagraphic images estimated from the deep-learning models are similar to the Solar Heliospheric Observatory (SOHO)/Large Angle and Spectroscopic Coronagraph C3 coronagraph data, especially during the solar minimum period. The AI-generated coronal density distribution from this study can be used for space weather models on a near-real-time basis.

Near-real-time 3D Reconstruction of the Solar Coronal Parameters Based on the Magnetohydrodynamic Algorithm outside a Sphere Using Deep Learning

“…We took the data regarding these parameters from the PSI website. 4 Following our previous work (Rahman et al 2023), we incorporate data augmentation techniques to enhance the amount of data. These provide diversity into the data set during the training process and help to address potential artifact issues during testing (Shorten & Khoshgoftaar 2019).…”

“…The size of our input and target data sets are 182 × 101 (longitude, latitude) and 182 × 101 × 150 (longitude, latitude, and altitude from the solar surface), respectively. We also normalize the data sets according to Rahman et al (2023). For analyzing the results of the test data sets, we denormalized and converted the MAS code units to cgs units, following the process described in Mikić et al (2018).…”

“…The Pix2Pix (Isola et al 2017) and the Pix2PixHD (Wang et al 2018) are supervised models based on a GAN. Those models have been applied in various areas of astronomy and space weather applications: (1) image translation among mutiwavelength images (Kim et al 2019;Park et al 2019;Jeong et al 2020;Shin et al 2020;Jeong et al 2022;Lawrance et al 2022), (2) denoising the solar magnetogram (Park et al 2020), (3) superresolution of the solar data (Jia et al 2019;Rahman et al 2020), and (4) the 3D reconstruction of coronal electron density (Jang et al 2021;Rahman et al 2023).…”

“…In our previous paper (Rahman et al 2023), we successfully produced a 3D density structure of the solar corona from 1 to 30 solar radii using the Pix2PixHD model (Wang et al 2018) whose computing time was much shorter than the MAS simulation. The synthetic coronagraphic data estimated from the results are consistent with the real coronagraph image of the Solar Heliospheric Observatory/Large Angle Spectroscopic Coronagraph C3 coronagraph data.…”

“…The synthetic coronagraphic data estimated from the results are consistent with the real coronagraph image of the Solar Heliospheric Observatory/Large Angle Spectroscopic Coronagraph C3 coronagraph data. Following Rahman et al (2023), we generate solar coronal parameters (magnetic field, radial velocity, and temperature) at a near-real-time basis using the Pix2PixCC model. Along with the density results, the AIgenerated 3D structure of the solar coronal magnetic field, radial velocity, and temperature from our Pix2PixCC models can be employed for the inner boundary conditions for heliospheric simulation models such as the MAS heliospheric models (Lionello et al 2009) and the Wang-Sheeley-Arge model (Arge et al 2004).…”